To render simulation more realistic, noise sounds generated during operating conditions are included in the simulation as simulated noise sounds. Simulated noise sounds are typically generated through filtering signals (e.g. signals captured during operating conditions), the isolation of various sources and a playback of the sources. Sound models outputs are mixed and sent to a distribution mixer. For example in the context of a flight simulator, sound models are generally further adapted with equalization filters. However, those filters do not take into account multiple parameters, including the simulator ambient noise, inter-model coherent signals, channel effectiveness, channels traveling paths, etc. Consequently, the equalization filters are calibrated manually by the factory before shipping, to try to take into consideration these parameters.
However, the manual factory calibration information may be lost when multiple filters are applied on sound models. An additional step is needed to adapt the sound models to meet a required global sound level. If a change in the simulator ambient noise occurs, the global sound level is affected which requires new adjustments to be performed on the sound models. The application of multiple filters also affects the traceability of the sound models with the initial raw data, making updates and new adjustments difficult to perform.
Another issue occurs when the quality of a plurality of loudspeakers used for playing the plurality of simulated noise sounds during simulation is not constant. In particular, the low frequency response of the loudspeakers may vary significantly, based on the cost and quality of each one of the loudspeakers used for the simulation. Consequently, the rendering of sound models having low frequencies is significantly affected by the low frequency response of the loudspeakers in charge of playing these sound models with low frequencies.
There is therefore a need for a new system for generating calibrated multi-channel non-coherent signals.